PACIFIC HERRING 
299 
can safely be assumed that spawning escapements are in the main roughly proportional to the 
catch, it becomes apparent that they are the predominating factor in determining the size of the 
runs. 
In the salmon, which spawn but once in a lifetime, such a correlation between 
spawners and offspring is easier to trace than with a fish that may spawn several 
times during its life. However, Jensen (1927) has attempted to trace just such a 
correlation in the case of the South Baltic autumn herring, a single local race of her- 
ring around Bornholm, spawning in September and October. He found that the 
catch anomalies (per cent from normal) of the herring fishery at Bornholm showed 
fluctuations covering three or four years. During 26 years he noted seven fluctuations 
averaging 3.7 years. Jensen says: 
The periodicity is to be explained as a consequence of the varying amount of fry produced by 
the varying number of spawning herrings. The Baltic herring generally matures in its third year 
and the shoals of spawning herrings accordingly consist of fish 3 years old and more. As the her- 
rings of the year classes III and IV in the Baltic greatly 
predominate over the older year classes, and therefore as 
a rule produce the largest quantity of fry, the observed 
periodicity can be explained in this way. 
V 
I 
0 
i* 
e> 
5 
§ 
1 
1926 
/927 
/ 2 3 4 
5 6 7 
AGL 
a 9 /O / / 12 
On the other hand the great fishery for the 
Atlanto-Scandian spring herring shows violent 
fluctuations due to dominant year classes that ap- 
pear at irregular intervals (Lea, 1924), seemingly 
without regard to the number of spawners present. 
Even in the case of a clear-cut periodicity 
as in the Karluk salmon, or of a lesser correla- 
tion as in the salmon of the Gudenaa (Jacobsen 
and Johansen, 1921, p. 12-15), it would seem that, 
owing to the fact that the salmon do not all return 
to their native stream the same number of years 
after leaving it (in the Karluk about 85 per cent 
return at 5 years of age and 10 per cent at 6 
years of age), in a comparatively limited number 
of years the effect of the spawning of a dominant year class would be so distrib- 
uted that soon its successors in the 5-year cycle would no longer be dominant. 
That such is not the case causes one to ask if there may not be some other condition 
present, possibly meteorological, that favors such a 5-year cycle. 
If the dominant year classes in the Alaska herring recur in regular cycles, as in 
the Bornholm autumn herring, then the trend of the fishery should not be hard to 
establish; but if they recur at irregular intervals, and if successive dominant year 
classes are of greater or lesser dominance, as in the Norwegian spring herring, then 
the solution may be more difficult. 
Figure 37.— Age histograms of Elrington Pas- 
sage, Prince William Sound, for 1925, 1926, 
and 1927 
EFFECT OF DOMINANT YEAR CLASSES ON THE CATCH 
The presence of dominant age groups may have a far-reaching effect; at times a 
race may be exceedingly abundant and at other times exceedingly scarce, for there 
may be periods of several years between dominant year classes, the population 
becoming much reduced before another dominant year class appears in the catch. 
The appearance of such a year class may cause excessive abundance for a time. 
When a very dominant year class first enters the commercial catch its members will 
